A lot of people are asking me what’s better right now: Claude Code or ChatGPT Here’s the answer: The gap in usage limits for the frontier models - Opus 5.5 and Sol 6.1 - between Anthropic and ChatGPT is truly massive The gap is much smaller when comparing Astra and Claude 5.1, though For the end user, it doesn’t really matter. We get the best of both worlds Source : newsletter.semianalysis.com/…

Hey @commonsmade, vouch @Mikadzyki_NFT

Boris Cherny - the creator of Claude Code > born in Ukraine (Odessa) in the early 90s1995 > at 4–5 years old emigrates to the US with his family > grandfather - one of the first programmers in the USSR, worked with punch cards > stacks of them all over the houseenters UCSD > studies economics instead of computer science > drops out after two years > launches his first startup at 18 > completely self-taught, no CS degree > goes through several startups, including Y Combinator > works in tiny teams where one person does > everythingmid 2010s > joins a hedge fund as an engineer > falls in love with functional programming and TypeScript > starts running TypeScript meetups in San Francisco2017 > joins Meta (Facebook) as Senior Software Engineer > grows quickly inside Instagram > becomes Frontend Architect and Lead modernizes massive codebases, migrates from Python to Hack/Comet > leads major GraphQL migrations > works on Messenger + Facebook Groups integration2019 > writes the book Programming TypeScript (O’Reilly) - one of the main books on the languageSeptember 2024 > leaves Meta for Anthropic > joins as regular Member of Technical Staff > in the first few weeks builds a simple CLI prototype - just because he wanted to write code with AI himself > this prototype turns into Claude Code > builds the cleanest and most minimal architecture > now writes 80–100% of his own code using Claude Code > becomes Head of Claude Code > Claude Code grows rapidly into a key product: - generates over $2.5 billion in annualized revenue for Anthropic - becomes one of the main drivers of the company’s growth February 2026 > @AnthropicAI raises $30 billion in Series G > valuation reaches $380 billion post-money > lives in San Francisco, married > continues launching personal side quests for growth > believes the best way to learn programming is by building real products > convinced that traditional IDEs like VS Code will change dramatically in the coming years Claude Code legend - @bcherny

Looking at what AI can already do today, it seems to me that soon they’ll be able to vibe-code an entirely new Polymarket for you This is just a side note about the kind of opportunities that vibe coding and skilled hands already open up right now

Does arbitrage exist on Polymarket? Today I’m going to check this A few days ago I found a free arbitrage bot on GitHub with open-source code Its working logic is as follows: BTC goes up (UP): $0.48 BTC goes down (DOWN): $0.51 ───────────────────────── Total: $0.99 ✅ < $1.00 Profit: $0.01 per share (1.01%) Why does it work? > At close, ONE of the two sides pays $1.00 per share. > If you paid $0.99 total, you earn $0.01 no matter which side wins. > It’s guaranteed profit (pure arbitrage). Personally, I have doubts about the very possibility of such a situation appearing on Polymarket, so I will most likely try to vibe-code my own trading bot using Claude or Cursor I’ll let you know what comes out of this bot and I’ll also keep you posted on the development of my own bot

Polymarket isn't about gambling Polymarket is about strict mathematical calculation and a cool head Arbitrage, find market inefficiencies - and you'll always end up in profit

While everyone is talking about arbitrage within Polymarket, everyone forgets about third-party web3 betting platforms Our team has launched a fully functional and ready-to-use parser that tracks arbitrage opportunities If you're interested in getting beta access comm this post

Lighter points now = $ZEC mid-pump

Lighter joins the rare ranks of crypto exchanges offering spot trading for major assets Currently available for spot trading: $BTC, $ETH, $SOL 3 reasons why this matters: / Instant hedging – neutralize perp risk with spot without moving funds / Funding rate arbitrage – spot + opposite perp = near-risk-free yield / PerpDex - Dex / Cex arbitrage This feature will soon move from testnet to @Lighter_xyz mainnet

I don't understand why you're still not trading on Lighter, ignoring the fact that you can earn points just for trading

Lighter FDV odds on Polymarket Current Polymarket probabilities look strange: - $2B+ FDV → 60% - $4B+ FDV → 26% Feels super weird because points are getting SNAPPED UP on OTC right now at $55–70 each The whole thing is tanking because of one killer condition: > If Lighter doesn't launch a token by December 31, 2025, 11:59 PM ET, this market will resolve to "No" That’s why the odds are collapsing - the market can just resolve NO if they miss the deadline Everyone is hedging on No and scared to go Yes, because in crypto TGEs get delayed forever while teams “wait for perfect conditions” Kalshi has much friendlier rules - deadline is the actual TGE date, no hard cut-off My take: I still expect ~$4B FDV (or higher) Compared to Aster MC $2.1B | FDV $8.32B, Lighter can easily aim for MC $1.2B+ | FDV $4B+ $100 per point one day

New @Lighter_xyz users milestone Total Users: 400,305 Milestone: 400K users Are we tired of winning yet?

I've seen this somewhere before...

. @Lighter_xyz new listings detected > $FIL > $ICP > $STRK Please note: these markets are likely not tradable yet - wait for an official announcement

Why People Still Choose Centralized Exchanges [Quotes and insights from the @ProMint_X AMA with @vnovakovski , CEO of Lighter Exchange] ⤷ 1. Old habits die hard “There’s inertia. People who bought Bitcoin 10 years ago on CEXs just keep using what they’re used to.” > Vladimir says it’s not about logic - it’s about comfort. People simply don’t like changing what feels familiar ⤷ 2.Self-custody still scares many “For some, it’s scary to hold their own keys. They feel safer having a password, something recoverable if they lose access.” > Most users aren’t afraid of DeFi - they’re afraid of losing their seed phrase or messing something up ⤷ 3. Awareness hasn’t caught up with technology “There are onboarding tools like Dynamic and Privy that let you sign up with just an email and password. But a lot of people don’t even know that yet.” > The tools are ready. The mindset isn’t. ⤷ 4. The risk is misunderstood “People think DeFi is riskier than it is. But in reality, centralized systems are the risky ones — they’re black boxes.” > He’s blunt here: CEXs feel safe, but they’re opaque. Users trust what they can’t verify ⤷ 5. Change is happening, just slowly “Two years ago DEX users were 1–2%, last year 5%, now over 10%. The trend’s going in the right direction, just not fast enough.” The shift is real - but we’re still early [The real risk isn’t in DeFi. It’s in centralized systems]

. @Polymarket gives Democrats a 71% shot at taking the House - solid, given the historic midterm headwind against the president’s party Little-known fact: Dem Jared Golden just announced he’s out; it’s the most Trump-friendly district held by any Democrat, so the GOP will almost certainly flip it If Trump’s approval stays below 50%, history since 1934 says the president’s party loses ~28 seats on avg - Dems need just +3 for a majority So I’m leaning Democrats

Today, the @Lighter_xyz community received a total of 250,000 points Here are some highlights from the past week: - **0x1fD** — top market maker - **0x41b** — top market maker - **0xb8C** — top market maker - **0xc5c** — largest liquidation - **0x4F2** — biggest loss Artificial trading activity does *not* earn points

GM Today I’ll be watching arbitrage opportunities between Polymarket, Kalshi, and Opinion By the way, the Opinion site lags terribly and is extremely user-unfriendly > From my experience, such projects tend to give out the best rewards

Gm CT I'm eligible for Stage 3 with the Bronze Early badge ► Mint start: 11/11 | price: 2.5 $SOL Total NFT supply - 1999: <> Stage 1 - 800 (3:00–6:59 PM) <> Stage 2 - 600 (7:00–7:59 PM) <> Stage 3 - 300 (8:00–8:59 PM) <> Stage 4 - 299 (9:00–9:59 PM) ► My thoughts: I don’t think the price will moon right after mint like Cerebro Keys did. The utility is pretty weak I’m leaning toward a quick flip after mint and then buying back closer to the snapshot to qualify for the @wallchain airdrop in Q4 TGE What are you planning to do?